Compare the two reviewers' answers to each RoB 2 signalling question for the same trial result, then resolve only the domains where those answers diverge. RoB 2 assesses a specific randomized-trial result across bias domains using signalling questions and written justifications, producing low risk, some concerns, or high risk judgements. The signalling questions are the shared unit of comparison, so the reconciliation note should record where the underlying answers differ and which evidence settled each difference.
- Fix the unit of assessment. Confirm both reviewers judged the same trial result, since RoB 2 applies to a specific result rather than to a trial as a whole. If one reviewer assessed a different outcome or time point, align the target before comparing judgements.
- Tabulate each reviewer's answer to every signalling question in the domain where the overall judgements differ. Mark each question as agree, diverge, or answered by only one reviewer. This isolates the questions that actually drive the disagreement.
- For each divergent signalling question, record the evidence each reviewer cited and the evidence that resolves it, such as the trial report section, protocol, or supplementary file. Write the resolved answer and the evidence that settled it in the note.
- Re-derive the domain judgement from the resolved signalling answers, then record the final overall judgement for that result. Keep the original two judgements alongside the resolved one so the trail from disagreement to decision stays visible.
CONSORT 2025 supports complete reporting of randomized trials; it is a reporting guideline and checklist, and it is not a study design or quality appraisal tool. Use CONSORT to check whether missing reporting explains a divergent answer, then return to the RoB 2 signalling question to record the resolved judgement.